Vectorizing C compilers: how good are they?

The programming language C is becoming more and more popular among users of high-performance vector computer architectures. With this popularity of C, it becomes more critical to have a good optimizing/vectorizing C compiler. This paper describes a study of four such vectorizing C compilers, with emphasis on the automatic vectorization ability of each compiler. This study is similar to the Fortran study that was described in [CDL88] and in fact, one facet of this study is a C version of the same kernels. Three suites of C loop kernels have been developed to determine the strengths and weaknesses of vectorizing compilers. The Convex cc compiler, the Convex Application Compiler, the Cray 2 sec compiler, and the Cray YMP sec compiler have been tested against these suites. The paper gives the results for each suite, with identification of problem areas for each compiler.